Remotely piloted aircraft system (RPAS) applied to the forest Inventory of Tectona grandis Linn f.: an RGB-based approach

Abstract

Advances in the forestry sector are closely linked to scientific research and the implementation of technologies in the areas of sensors, aerial images, GPS, learning algorithms, data processing, as well as the use of computer intelligence systems to obtain and analyze data in forestry plantations, which require tools that provide a higher degree of precision, reduced costs and less time for data collection. This constant evolution of the sector, together with the improvement of Remotely Piloted Aircraft Systems (RPAS), paves the way for the use of this platform in forest data collection. In this article we discussed the use of RPAS to obtain dendrometric and morphometric variables to estimate wood volume in two Tectona grandis Linn f. stands, where field measurement and overflight were carried out using the Phantom 4 Pro RPAS. The aerial photos were processed using Pix4Dmapper software to generate the orthomosaic, digital terrain model and surface. Using these products, information was collected on the total height, crown area and crown diameter of the trees, with the aim of estimating volume (v) using the variables collected from the images (indirect method) and with the addition of the variables collected in the field (direct method). The estimates of v showed satisfactory results through the evaluation statistics, and it was possible to estimate forest production from data obtained by RPAS, obtaining an adjusted coefficient of determination of 66% in the indirect method and the estimate together with the field data showed an adjusted coefficient of determination of 94%.